WO2014088490A1

WO2014088490A1 - Control of at least one state in a system toward a reference value

Info

Publication number: WO2014088490A1
Application number: PCT/SE2013/051411
Authority: WO
Inventors: Robin ROCKSTRÖM; Martin Evaldsson; Elvedin Ramic; Frank Mohr
Original assignee: Scania Cv Ab
Priority date: 2012-12-04
Filing date: 2013-11-29
Publication date: 2014-06-12
Also published as: SE1251365A1

Abstract

A method and a system for the control of at least one state in a system toward a reference value is presented, where said at least one state comprises an inertia and said control is carried out with the use of a control signal. According to the present invention said control is model-based, where the model comprises a force equation related to said system. Further, a size of said control signal is proportionate to a change in said at least one state.

Description

CONTROL OF AT LEAST ONE STATE IN A SYSTEM TOWARD A REFERENCE

VALUE

Technical field

The present invention pertains to a method for the control of a state in a system according to the preamble of clam 1. The present invention also pertains to a control system arranged for the control of a state in a system according to the preamble to claim 27 and a computer program and a computer program product, which implement the method according to the invention . Background

Automatic control systems comprising one or several regulators are currently used for the control of a large number of various types of systems. The control often comprises that a state is controlled toward a reference value for the state. Many of the systems to be controlled by such control systems have an inertia, for example a mass inertia, a thermal inertia or a moment of inertia. Inertia in this context and in this document means a resistance to change, for example to a change in movement or a change in temperature, which means that changes do not occur momentarily, but the change occurs over a period of time.

There are several types of regulators according to prior art. Below is a description of the function and algorithm of a PID regulator, but other types [ types ] /variants of regulators may also function in a similar manner.

A PID regulator is a regulator which provides an input signal u(t) to a system that is to be controlled based on a

discrepancy e(t) between a desired output signal r(t), i.e. the reference value, and a real output signal y(t) . Below, e(t) = r(t) - y(t) according to the following:

where :

- K_p constitutes a gain constant;

- Ki constitutes an integration constant; and - K_D constitutes a derivation constant.

A PID regulator regulates in three manners, by way of a proportional gain (P; K_p) , by way of an integration (I; Ki) , and by way of a derivation (D; K_d).

The constants K_p, Ki and K_d impact the system as follows. An increased value for the gain constant K_p leads to the following change of the PID regulator:

- increased speed;

- reduced stabilisation margins;

- improved compensation of process disturbances; and - increased control signal activity.

An increased value for the integration constant Ki leads to the following change of the PID regulator:

- better compensation of low-frequency process disturbances (eliminates remaining faults in step disturbances); - increased speed; and

- reduced stabilisation margins An increased value for the derivation constant K_d leads to the following change of the PID regulator:

- increased speed

- increased stability margins; and - increased control signal activity.

As mentioned above, there are also other types/variants of regulators/regulation algorithms which have a function similar to that of the PID regulator.

Brief description of the invention

Previously known regulators/regulation algorithms, such as the PID regulator, have a large number of parameters which need to be calibrated in order for the control of the systems which they are designed to control to be exact and accurate. This means that calibration often needs to be carried out by somebody who has good knowledge of the system to be controlled and who also has a good understanding of the regulator

function. Unfortunately, the person who owns or uses the system seldom has the detailed knowledge about the system and/or regulator algorithm which is needed for calibration to provide an optimal control system. The fact that for example the PID regulator is so commonly used in regulator systems is partly due to the fact that it is relatively robust against incorrect calibration. Thus, the PID regulator often provides a functioning control system. However, this control system is rarely optimal for the control of the system which it is intended to control. In other words, the PID regulator

provides a regulation which functions for most systems, however, the PID regulation is not particularly exact. Thus the PID regulation rarely follows the desired reference values particularly well, which often leads to substandard performance and/or unnecessary costs.

For example, a PID regulation in a cruise control system of a vehicle adjusts relatively poorly to a reference speed which the cruise control has calculated that the vehicle should maintain in order to minimise fuel consumption, which leads to unnecessarily high fuel consumption in the vehicle. The same applies to other known regulators with similar functions to that of the PID regulator. One problem with the PID regulator, which is also a

contributing cause for the inaccuracy of the PID regulation, is that the PID algorithm gives rise to over- and/or

undershoots for the control signal to the system which is controlled, i.e. the input signal u(t) . These over- and/or undershoots are due at least partly to a winding up in the regulator caused by the gain from the I-part, i.e. from the integrating part (I; ¾) of the algorithm. The winding up functions like a memory that remembers too far back in time, which means there is a risk that the regulator may compensate for old faults. There are also similar problems for other regulators that comprise an at least partly integrating function .

One objective of the present invention is to provide a control system which resolves the above-mentioned problem with prior art control systems.

This objective is achieved through the above-mentioned method in accordance with the characterising portion of patent claim 1. The objective is also achieved through the above-mentioned control system according to the characterising portion of patent claim 27, and the above-mentioned computer program and computer program product. Through the present invention, at least one state in a system is controlled toward a reference value, where the at least one state has an inertia, for example a mass inertia, a thermal inertia or a moment of inertia, as explained above. The control according to the invention is based on a model

comprising a force equation which is related to the system to be regulated. The control of the at least one state is carried out with the use of a control signal, which according to the invention shall be proportionate to the change, i.e. the derivative of the at least one state that is controlled.

Thus the control of the states according to the present invention is based on the systems which comprise the

respective states. In other words, the systems are regulated based on the systems themselves, since the regulation of the systems is carried out based on models of the systems to be regulated. Thus the knowledge of the systems to be regulated within the regulation algorithm is good, which means that the regulation algorithm may be made stable.

Since the regulation according to the invention is based on the system to be regulated, a regulation as powerful or delicate a regulation as the system requires is obtained. Thus an adaptive adjustment of the regulation is obtained, where the adjustment follows the change of the system. This means that a very accurate regulation of the system may be obtained when the present invention is used.

The regulation according to the invention does not need to use predetermined gain factors as required in e.g. PID regulation, which means that the regulation may be easily adapted to the system which is to be controlled. In addition, the number of parameters that must be calibrated is radically reduced with the present invention compared to e.g. a PID regulator. Since a physical model for the system to be regulated is the basis of the regulation, considerably less knowledge within regulation technology is required in order to tune the control system. Essentially no detailed knowledge in regulation technology or regulation algorithms is required in order to handle the control system according to the present invention.

The regulation according to the present invention adapts quickly and well to the reference value since over-- and/or undershoots in the control signal, which have been problematic for prior art regulation algorithms, do not arise in the regulation according to the present invention. In particular, the winding up in prior art regulators due to gains from the I-part may be avoided with the invention.

According to one embodiment of the invention, the system to be regulated is a cruise control system in a vehicle, which has an inertia that is related to a mass m for the vehicle. The state here consists of an actual speed v_act for the vehicle. Where the present invention is used in such a cruise control system, the vehicle's fuel consumption may be reduced since the regulation follows the reference speed determined by the cruise control system more accurately.

According to one embodiment of the invention, the system to be regulated is an engine in a vehicle, with an inertia based on a moment of inertia / for the engine. The at least one state here comprises an actual speed <j)_act for the engine. Through regulation according to the present invention, the speed of the engine may be controlled very accurately without any extensive calibration work.

According to one embodiment of the invention, the system to be regulated is a temperature control system, with an inertia based on a thermal inertia K. The at least one state here comprises an actual temperature T_act . Thus a very accurate adjustment of the actual temperature T_act to the desired

reference value is achieved, which in certain environments, e.g. laboratories, is decisive to the activity and in other environments, e.g. office premises and vehicle cabins,

increases the comfort level for the staff, for instance.

According to one embodiment of the invention, the system to be regulated is a system for acceleration limitation in a

vehicle, which has an inertia based on a mass m for the vehicle. The at least one state here comprises an actual acceleration a_act for the vehicle. Thus a very good control over the vehicle acceleration may be obtained.

According to one embodiment of the invention, the system to be regulated is a system for braking of the vehicle, which has an inertia based on the vehicle mass m. The at least one state here comprises an actual speed v_act for the vehicle. Thus, essentially all types of braking systems for a vehicle

suitable for this purpose may be controlled very accurately. Brief description of figures

The invention will be illustrated in more detail below, along with the enclosed drawings where similar references are used for similar parts, and where:

Figure 1 shows an example of a transient toward a reference value,

Figure 2 shows an example of a transient toward a reference value,

Figures 3a and 3b show examples of a determination of a vehicle's mass, and Figure 4 shows a control device. Description of preferred embodiments

The present invention pertains, according to one aspect, to a method for the control of one or several states in a system toward a reference value and according to one aspect toward a control system arranged for the control of this at least one state toward the reference value. States for the systems that can be regulated according to the present invention have an inertia, which means that there is a resistance to change in relation to the state, for example to a change in movement or a change in temperature. Changes of states therefore occur over a time period and not momentarily.

Many systems have an in-built inertia for their states.

Essentially all such systems may be regulated with the use of the present invention. A couple of embodiments of the present invention will be described below. However, a person skilled in the art will realise that the present invention may

generally be applied in essentially all systems where the systems' states have some sort of resistance to change, i.e. some type of inertia.

For example, the temperature in a temperature control system may be controlled toward a temperature reference value with the present invention. In an engine control system in a vehicle, the engine speed may be controlled toward a reference speed. In a cruise control system of a vehicle, the vehicle speed may be controlled toward a reference speed. In a brake control system in a vehicle, the engine speed may be

controlled toward a reference value in the form of a maximum speed. In a control system for acceleration limitation, the vehicle acceleration may be controlled toward a reference acceleration. These embodiments and applications of the invention will be described in further detail below.

The control of the at least one state in the system to be regulated is, according to the invention, based on a model. The model which is used here is related to the system to be regulated in that the model comprises a force equation or another equation which is related to this system. Thus, a model of the system is produced at least partly in that a force equation is set up for at least a part of the system. Further, when the system is regulated, the control of the state is achieved with the use of a control signal, where the model-based control entails that the size of this control signal is proportionate to the change in the at least one state. Thus, a control device provides the control signal based on the model so that its size is proportionate to a change in the state.

Regulating systems based on appearance and characteristics of the systems themselves makes the regulation of the systems very accurate. As opposed to prior art control systems, in which the systems to be regulated have been forced into the regulation algorithms of the control systems, e.g. in the form of gain parameters, the regulation algorithm according to the present invention is based entirely on the system to be regulated, so that an optimally adapted regulation may be provided when the present invention is used.

The number of parameters to be calibrated in the control system before the regulation is reduced radically with the present invention compared to prior art regulators. This means that considerably less knowledge in regulation technology is required to tune the control system so that a good regulation is achieved. In addition, the regulation according to the present invention minimises over- and/or undershoots in the control signal, which arose in prior art regulators due to the integrating memory part of its regulation algorithms, i.e. in the I-part of a PID regulator, for example.

According to one embodiment of the invention, the system to be regulated is a cruise control system in a vehicle, which has an inertia related to a mass m for the vehicle. The state to be controlled here constitutes an actual speed v_act for the vehicle, i.e. the actual speed which the vehicle will maintain as a result of the cruise control. The reference value toward which the state is controlled here constitutes a reference speed v_ref for the vehicle.

There is a number of different types of cruise controls for vehicles. In some of these types of cruise controls the driver sets the reference speed v_ref . In other types of cruise

controls, the driver sets a set speed v_set , based on which the cruise control then determines the size of the reference speed v_rej sent to the speed regulator, so that the reference speed ^vref may have another value than the set speed v_set .

According to the embodiment, the model takes into account the difference between an actual acceleration a_act of the vehicle, i.e. the actual acceleration which is a result of the cruise control, and a reference acceleration _ref for the vehicle.

This difference is due to a time parameter τ, which is

described in further detail below. The time parameter τ

determines the appearance of the transient for the actual speed v_act toward the reference speed v_rej , so that a smaller value of the time parameter τ produces a quick transient and a larger value of the time parameter τ produces a slow transient. This is showed schematically in Figure 1, where the dashed straight horizontal line is a reference speed v_act against which actual speeds for different values of τ turn. As illustrated in Figure 1, the smallest value of the time parameter τ=2 (solid curve) produces the quickest transient, the larger value of the time parameter τ = 5 (dotted curve) produces a slower transient, and the largest value of the time parameter τ = 8 (dashed curve) produces the slowest transient.

According to the embodiment, the time parameter τ thus

constitutes the only parameter to be calibrated for the regulation to achieve the desired course. In other words, the value for the time parameter τ determines how long it takes for the actual speed v_act to be equal to the reference speed v_re .

Thus, according to the invention, the time parameter τ should be adjusted to provide, for the system to be regulated, an optimal appearance of the transient for the actual speed v_act toward the reference speed v_ref . It is relatively simple and intuitive for a user of a control system according to the present invention to realise the impact of an adjustment of the time parameter's size τ on the regulation.

Thus only one parameter, the time parameter τ, needs to be calibrated here, and it is also easy to understand and explain the impact this parameter has on the regulation. This is a considerable simplification and improvement for the user compared to previous regulators, such as PID regulators, where a number of parameters, having an impact on the regulation which is difficult to understand for the user, must be

adjusted to adapt the regulation to the system to be

controlled. For example, parameters for the P-part, the I-part and the D-part must be calibrated when a PID regulator is adjusted to a system which it is to control. The impact of these adjustments of the P-, I-- and D-parts on the regulation is at best incalculable for the user, and at worst totally incomprehensible . The present invention therefore presents a considerable

simplification of the calibration which is required to carry out the regulation of the system.

According to one embodiment of the present invention, the value of the time parameter τ is related to a run mode, also called a driving mode, for the vehicle. This is shown

schematically in Figure 2. The value of the time parameter τ is seen here as related to an aggressivity of the regulation.

Thus, in a normal driving mode, e.g. called "standard" (dotted curve), the time parameter τ may be given a medium-sized value. For a more aggressive driving mode, e.g. called "power"

(dashed curve) , the time parameter τ is given a relatively small value if the actual speed v_act is lower than the reference speed v_ref . For this driving mode the time parameter τ is given a relatively large value if the actual speed v_act is higher than the reference speed v_ref , as shown in Figure 2. The more

aggressive driving mode "power" thus turns inwards quickly when it approximates the reference speed v_ref from below (from a lower speed) , but turns inwards slowly when it approximates the reference speed v_ref from above (from a higher speed) . The driving mode "power" thus attempts to achieve the reference speed quickly v_ref from below and meets up early from above, which produces a powerful impression, a higher medium speed and a time gain compared to the other driving modes.

For a less aggressive driving mode, e.g. called "eco" (solid curve) , the time parameter τ is given a relatively large value if the actual speed v_act is lower than the reference speed v_rej . In the same manner the time parameter τ for this driving mode is given a relatively small value if the actual speed v_act is higher than the reference speed v_ref, as shown in Figure 2. The less aggressive driving mode "eco" thus turns inwards quickly when it approximates the reference speed v_ref from above (from a higher speed) , but turns inwards slowly when it approximates the reference speed v_ref from below (from a lower speed), which gives a soft impression, and a lower average speed and thus a lower total fuel consumption. In addition, the amount of braked away energy is reduced with the driving mode "eco", since a vehicle e.g. on a road section comprising an uphill slope followed by a downhill slope has a lower speed at the top of the hill. Due to the lower speed at the top, less braking is required downhill, so that less energy is braked away .

As shown in the non-limiting example in Figure 2, the

regulation according to the present invention can easily be adjusted so that different driving modes for the vehicle are obtained. The only parameter which needs to be changed to achieve this is the time parameter τ, which is a considerable simplification compared to prior art control systems.

According to one embodiment where the invention is applied in cruise control, the force equation which the model for the system to be regulated describes is related to an appearance as follows:

F_k+1 = m * ^^ref ^~Vact - a_ac^j + F_{k r} where (Equ.l)

- F_k+1 is a force which will impact said vehicle at the next iteration of the equation is calculated;

- m is the mass of the vehicle; - v_rej is the reference speed;

- v_act is the actual speed;

- I is the time parameter;

^{~~ a}act is the actual acceleration for the vehicle; and - F_k is a current force which impacts the vehicle.

^ref—^ ct

As shown m Equation 1, the term corresponds to a

T

reference acceleration a_ref , which means that the size of the difference for the time parameter τ has a direct impact on the difference between the reference acceleration _ref and the actual acceleration a_act . Thus, a simple adjustment of the value for the time parameter τ may impact the entire regulation according to the present invention, in the manner described above .

Since the present invention is applied to a cruise control system, the vehicle's fuel consumption may be reduced because the regulation follows the reference speed determined by the cruise control system more accurately. This is due to the fact that the regulation according to the invention is based on a physical model of the cruise control system, i.e. that the regulation is based on the system to be regulated.

The signal for the actual acceleration a_act which the vehicle may provide often comprises noise. This signal may be

generated in a number of different ways in the vehicle. For example, the acceleration may be determined by way of

derivation of the actual vehicle speed v_act, with the use of the force equation or with the use of an accelerometer. Regardless of how the signal for the actual acceleration a_act is generated, it is often noisy, which means it is difficult to base the regulation on it since there is a risk that it might be unstable .

According to one embodiment of the present invention, the force equation is therefore adjusted to compensate for the noise in the signal for the actual acceleration a_act . The adjusted force equation then looks as follows:

^„ h (Vref -Vact . _Γ , , _π

^Fk+i = -m * {—^. aact) + , where(Equ. 2)

- F_k+1 is a force which will impact the vehicle at the next iteration;

- h is a discretion factor, which specifies a resolution in the control system, where this resolution may be related to the sampling time/time step of the control system, which may e.g. be 0.01 seconds; - γ is a calibration time;

- m is the mass of the vehicle;

- v_rej is the reference speed;

- v_act is the actual speed;

- I is the time parameter; - a_act is the actual acceleration;

- F_k is a current force which impacts the vehicle.

The value for the calibration time γ specifies how quickly the actual acceleration a_act approximates the reference acceleration a_ref . In other words, the value for the calibration time

(V_ref—V_acf \

~ ^aact I = 0 in Equation 2. By using the compensated force equation in Equation 2, the oscillation problem may be avoided for the regulation where the signal for the actual acceleration a_act is noisy. Also problems related to erroneous estimations of the vehicle mass m in the model may be avoided through this embodiment of the invention .

According to this embodiment, the regulation of the system is made oscillation-free, i.e. the control of the state is made non-oscillatory, by giving the time parameter τ a value which is at least four times larger than the value for the

calibration time γ, τ≥4*γ. The τ≥4*γ transient for the actual value toward the reference value then occurs entirely without over-- and/or undershoots. Oscillations in the

transient itself for the actual value toward the reference value are thus then avoided τ≥4*γ.

According to another embodiment, the time parameter τ is given a value which is at least more than four times as large as the value for the calibration time γ, τ>4*γ. For example, the time parameter τ is here given the value 5_*y, τ = 5*γ,

providing an additional 20% stability margin compared to the value T = 4_*y. Higher values for the time parameter τ may also be used, for example τ = 6*γ, or τ = 7 * γ, which provides

additional margins toward oscillation in the regulation. The higher values for the time parameter τ may be used to provide additional margins toward erroneous estimations of the vehicle mass m.

According to one embodiment of the present invention, the system to be regulated is a system for acceleration limitation in a vehicle. The state to be controlled then constitutes an actual acceleration a_act for the vehicle and the reference value used in the control constitutes a reference acceleration a_re^ for the vehicle. The inertia in the system for acceleration limitation is here based on the vehicle mass m. The physical model, on which the regulation is based, takes into account the difference between the actual acceleration a_act and the reference acceleration o,_ref , where the force equation looks as follows: F_k+1 = m * {o._ref — ⁽½ct) + F_k, where (Equ. 3)

- F_k+1 is the force which will impact the vehicle at the next iteration; - m is the mass of the vehicle;

- a_ref is the reference acceleration;

- a_act is the actual acceleration; and

- F_k is a previous current force which impacts the vehicle.

Thus, through this embodiment of the invention, the actual acceleration is controlled _act toward the reference

acceleration _rej so that the actual acceleration a_act is limited. Since the regulation is based on a physical model of the acceleration limitation system, thus an accurate control of the actual acceleration is obtained a_act r which is also easy to calibrate. The size of the time parameter τ determines in the manner described above how the transient appears when the actual acceleration a_act approximates the reference acceleration CL_ref , so that different values of the time parameter τ produce different behaviours for the acceleration limitation system. According to one embodiment of the present invention, the system to be regulated consists of a system for braking the vehicle. Essentially any type of vehicle braking system may be regulated according to this embodiment, e.g. a brake, a retarder, or an electromagnetic brake, which may e.g. consist of an electric engine in a hybrid vehicle. The at least one state here consists of an actual vehicle speed v_act and the reference value consists of a maximum vehicle speed v_max, the value of which may e.g. be based on speed limits for a road section. The inertia for the vehicle braking system is based on the vehicle mass m.

For this embodiment of the invention, the model of the vehicle braking system takes into consideration a difference between the actual vehicle acceleration a_act and the reference

acceleration _ref for the vehicle where the force equation looks as follows: B_k+1 = m * {^^max-^Vact _ a_act^ + B_k, where (Equ. 4)

- B_k+1 is a force which will impact said vehicle at the next iteration of the algorithm; - m is the vehicle mass;

- v_rej is the reference speed;

- v_act is the actual speed;

- I is the time parameter;

- a_act is the actual vehicle acceleration; and

- B_k is the current braking force which impacts the vehicle.

As indicated by Equation 4, the difference between the actual vehicle acceleration a_act and the reference acceleration _ref depends on the time parameter τ , since the reference

acceleration a_rej corresponds to the term ^Vmax~Vact _# j_n the same manner as described above, the size of the time parameter τ determines the appearance of the transient when the actual vehicle speed v_act approximates the maximum vehicle speed v_max. In this way, the vehicle braking system may be easily

calibrated and may potentially be provided with an altered behaviour, only by changing the value of the time parameter τ, while a very accurate regulation is obtained at the same time. Also for this embodiment, the above-described adjustment of the force equation may be applied to the force equation in

h ,

Equation 4 by introducing the ratio - m Equation 4, where h. is a discretion factor and γ is a calibration time, so that problems related to noise in the actual acceleration a_act and/or erroneous estimations of the vehicle mass m may be minimised.

As mentioned above, the control system may be made

oscillation-free, i.e. non-oscillatory, by giving the time parameter τ a suitable value, which is a value that is at least four times larger than the value for the calibration time γ, τ≥4*γ. Even larger margins towards transients/oscillation may be obtained by using larger values for the time parameter τ , e.g. =5*y, τ = 6*γ, or τ = 7*γ.

According to one embodiment of the present invention, the system to be regulated consists of an engine in a vehicle. The state to be controlled with this regulation then consists of an actual engine speed <j⁾ _act for said engine, and the reference value toward which the actual engine speed <j⁾ _act should be controlled constitutes a reference engine speed d_ref for the engine. The inertia for the engine system here consists of a moment of inertia / for the engine. The model for the engine system, on which the regulation is based, here takes into consideration the difference between a change of an actual engine speed <j⁾ _a ^' _ct for the engine and the change of a reference engine speed w_re/ f°^r the engine. The difference here depends on a time parameter τ, since the term T comprises a time parameter τ, so that the course of the transient of the actual engine speed <j⁾ _act toward the reference engine speed d_ref may be controlled through the size of the time parameter τ, in the same manner as described above for different embodiments. According to the embodiment, the model's force equation looks as follows:

= / * {^^re{^^{~ act} - o)_act^ + T_k , where (Equ. 5)

- Tfc+i is ^a torque which will be emitted by the engine at the next iteration of the algorithm;

- / is the moment of inertia for the engine;

- <j)_ref is the reference engine speed for the engine; ^{~ ω}αα is the actual engine speed of the engine;

- I is the time parameter; - <j_a ^' _ct is a change of the actual engine speed; and

- T_k is the torque which is currently emitted by the engine.

An accurate and easily calibrated regulation of the engine system is obtained through the regulation algorithm in

Equation 5. According to one embodiment of the present invention, the system to be regulated consists of a temperature control system for a limited volume, where the temperature control system's inertia is based on a thermal inertia K for the volume. The at least one state here constitutes an actual temperature T_act for the limited volume and the actual

temperature T_act is controlled towards a reference temperature T_ref for the volume.

The model for the temperature control system takes into consideration the difference between a change of an actual temperature T_act for the volume and a change of a reference temperature T_ref for this volume. The difference here depends on the time parameter τ and an equation for the temperature regulation, which according to the model of the temperature control system appears as follows:

P_k+1 = K * (^Tref-^Tact - T_act) + P_k, where (Equ. 6) - P_fc+i is a thermal effect which will be emitted in the limited volume at the next iteration of the algorithm;

- K is the thermal inertia for the limited volume;

- T_ref is the reference temperature;

- T_act is the actual temperature;

τ is the time parameter;

- T_act is a change of the actual temperature; and

- P_fc is a current thermal effect which is emitted in the limited volume.

The aggressiveness of the transient for the actual temperature

Tact toward the reference temperature T_re may easily be cancelled by adjusting the value for the model's time

parameter τ, so that the transient's character is changed as described in detail above.

According to one embodiment of the present invention, the system to be regulated consists of any suitable system

connected to a PTO in the vehicle. Certain vehicles, such as trucks and tractors, have PTOs to which a user may connect any type of equipment, such as a crane, a concrete mixer, or various types of power units. In vehicles with prior art control systems implemented, the user has then been able to choose between a predetermined number of preset calibrations of the PID regulator for this PTO, which has not been

satisfactory for the users. The great variation among the different types of systems that may be connected into PTOs means that regulators with a relatively large amount of different characteristics are required in order to operate these systems in a satisfactory manner.

Where the present invention is used for the PTO, the

regulation of the system connected to the PTO may, in the manner described above, be given a desired

character/aggressiveness through a simple adjustment of the time parameter τ. Thus, the aggressiveness of the regulation my be chosen essentially freely by the user, so that it may be optimised in relation to the system that is to be regulated. Thus, a very well adjusted regulation of all the various suitable systems which may be connected to the PTO in the vehicle is obtained.

According to one embodiment, the time parameter is given τ a relatively large value when the regulation of systems

connected to the PTO occurs, e.g. τ = 4*γ, =5*7, τ = 6*γ, or τ = 7*γ, which gives security margins in order for the

regulation to remain non-oscillating.

The present invention may also be used to determine the vehicle mass m. This is done by way of analysis of the

transient for actual state values in relation to the reference value. Here, the appearance of the actual transient is

compared, e.g. for the actual vehicle speed v_act, with an expected appearance for the same transient, e.g. with an expected appearance for this vehicle speed v_actexpect . If these two transients differ, this may be due to an error in the estimation of the vehicle mass, which means that the

regulation according to the invention becomes somewhat

inaccurate, so that the actual vehicle speed v_act has another appearance than it should have. Therefore, the estimation of the vehicle mass m is adjusted based on this analysis of the transient. In order to estimate a ratio between the actual mass m of the vehicle and the estimated mass m* of the vehicle, a mathematical analysis of a transient is carried out, which is described below. The vehicle always follows a force equation (Newton's second law) :

m .3— F driv- Fambientk (Equ. 7)

Where the vehicle is actually controlled by the regulator, i.e. where the regulator gets what it requests and where the control system is not in saturation in relation to maximum torque and drag torque, the speed profile for the actual vehicle speed v_act will follow a predetermined profile which only depends on the two parameters τ and γ , which may be deduced as set out below.

The vehicle is controlled, when the regulator actually

controls, by the equation:

+ F_k , where (Equ. 8]

- F_k+1 is a force which will impact the vehicle at the next iteration;

- h is a discretion factor;

- γ is a calibration time;

- m is the mass of the vehicle;

- v_ref is the reference speed; - ^Vact,k is the actual speed;

- I is the time parameter; ^{~ a}act,k is the actual acceleration; and

- F_k is a current force which impacts the vehicle.

By combining the force equation (Equation 7) and the power updating equation (Equation 8), the following expression is obtained :

„ _ hm* fVref-Vact,k \ „

m^aact,k+l ^< r ambientk+1 ^~ ~γ~ τ ^aact,kJ ^< 'ambientk^{+ ma}act,k

(Equ. 9)

Let us assume that the ambient power ^^' _omgivning is constant from one sample to the next, which is a reasonable assumption in e.g. a cruise control system, which is relatively slow.

Following some algebraic reshuffling, the following expression is obtain

If a speed error is defined as ε = v_ref -v_act and uses that the

(Cl — Cl I

act,k act,k+\ / ·

term is the numerical derivative of acceleration,

h

the following ordinary differential equation of the second order is obtained instead of the speed error if a transition is also made from discrete time to continuous time:

0 = + έ, where

(Equ. 11) m _* μ=— is the ratio between the hitherto estimated mass m m

and the actual mass m From Equation 11 the mass ratio μ may be easily resolved and calculated. The problem is that both έ and έ are often very noisy, so that the estimation therefore often becomes noisy as well .

In order to minimise the problem of noise in measuring

signals, the equation is integrated as of when the regulator actually starts to control the vehicle at the time t=0 until the system stabilises around the reference after the time t=T . The following expression for the mass ratio is then obtained:

Equation 12 is easy to realise in a discrete control system and it is guaranteed to converge as long as the speed

converges toward the reference. When this value for the mass ratio μ is calculated according to Equation 12, a new mass estimation m^* may be obtained by multiplying the old estimation m with the calculated mass ratio μ :

This may be repeated at each transient toward the reference.

A non-limiting simulated example of mass estimation according to this embodiment is shown in Figures 3a and 3b, where the solid curve corresponds to a transient where the mass

estimation is correct and the dotted curve corresponds to a transient with an erroneous mass estimation. For the correct mass estimation, the transient becomes oscillation-free (shown in Figure 3a) and the mass ratio = 1 (shown in Figure 3b) .

For the incorrect mass estimation, the transient obtains an overshoot (shown in Figure 3a) and the mass ratio ju = 0.5 (shown in Figure 3b) . The mass is thus here underestimated by 50%.

Figure 3b shows clearly that the algorithm converges toward the mass ratio values = 1 respectively μ = 0.5 for a correct and incorrect mass estimation, respectively, which means that the algorithm becomes very useful for purposes of correcting incorrect mass estimations. Additionally, the algorithm may be implemented with very low added complexity.

A person skilled in the art will realise that the method according to the present invention may also be implemented in a computer program, which when executed in a computer will cause the computer to carry out the method. The computer program usually consists of a computer program product 403 stored on a digital storage medium, where the computer program is comprised in the computer program products' computer readable medium. Said computer readable medium consists of a suitable memory, for example: ROM (Read-Only Memory), PROM

(Programmable Read-Only Memory) , EPROM (Erasable PROM) , Flash, EEPROM (Electrically Erasable PROM), a hard disk device, etc.

Figure 4 shows schematically a control device 400. The control device 400 comprises a calculation device 401, which may consist of essentially a suitable type of processor or

microcomputer, e.g. a circuit for digital signal processing (Digital Signal Processor, DSP) , or a circuit with a

predetermined specific function (Application Specific

Integrated Circuit, ASIC) . The calculation device 401 is connected to a memory unit 402 installed in the control device 400, which provides the calculation device 401 with e.g. the stored program code and/or the stored data which the

calculation device 401 needs in order to be able to carry out calculations. The calculation device 401 is also set up to store interim- or final results of calculations in the memory device 402.

Further, the control device 400 is equipped with devices 411, 412, 413, 414 for receiving and sending of input and output signals,- respectively. These input and- output signals may comprise wave shapes, pulses, or other attributes, which may be detected by the devices 411, 413 for the receipt of input signals as information and may be converted into signals that may be processed by the calculation device 401. These signals are then provided to the calculation device 401. The devices 412, 414 for sending of output signals are arranged to

transform signals received from the calculation device 401 for the creation of output signals by e.g. modulating the signals, which may be transmitted to other parts of the control system and/or systems regulated according to the present invention. Each one of the connections to the devices for receipt and sending of input and- output signals may consist of one or several cables; a data bus, such as a CAN (Controller Area Network) bus, a MOST (Media Oriented Systems Transport) bus, or any other bus configuration; or of a wireless connection. A person skilled in the art will realise that the above- mentioned computer may consist of the calculation device 401 and that the above-mentioned memory may consist of the memory device 402.

Generally, control systems in modern vehicles consist of a communications bus system consisting of one or several

communications buses to connect a number of electronic control devices (ECUs), or controllers, and different components localised on the vehicle. Such a control system may comprise a large number of control devices, and the responsibility for a specific function may be distributed among more than one control device. Vehicles of the type shown thus often comprise significantly more control devices than as shown in Figure 4, which is well known to a person skilled in the art within the technology area.

The present invention is implemented in the embodiment

displayed in the control device 400. The invention may, however, also be implemented wholly or partly in one or several other control devices already existing in the vehicle or a control device dedicated to the present invention.

A person skilled in the art will also realise that the above system may be modified according to the different embodiments of the method according to the invention. In addition, the invention pertains to a motor vehicle, e.g. a truck or a bus, comprising at least one control system according to the invention . The present invention is not limited to the embodiments of the invention described above, but pertains to and comprises all embodiments within the protected scope of the enclosed

independent claims.

Claims

Patent claims

1. Method for the control of at least one state in a system toward a reference value, where said at least one state comprises an inertia and said control is carried out with the use of a control signal characterised by the fact that

- said control is model-based, where the model comprises a force equation, or another equation related to said system; and

- a size of said control signal is proportionate to a change in said at least one state.

2. Method according to claim 1, where said system is a cruise control system in a vehicle, said at least one state consists of an actual speed v_act for said vehicle, said reference value consists of a reference speed v_ref for said vehicle, and said inertia is related to a mass m for said vehicle .

3. Method according to claim 2, where said model takes into consideration a difference between an actual acceleration ^aact f°^r said vehicle and a reference acceleration _ref for said vehicle, where said difference depends on a time parameter τ .

4. Method according to any of claims 2-3, where said force equation appears as follows:

fV_ref—V_acf \

F_k+1 = m * a_act)+F_k, where

- F_k+1 is a force which will impact said vehicle at the next iteration;

- m is said mass of said vehicle; - v_ref is said reference speed;

v_act is said actual speed; τ is a time parameter;

^{~ a}act is ^an actual acceleration for said vehicle; and

- F_k is a current force which impacts said vehicle.

5. Method according to claim 2, where said force

equation is adjusted to compensate for the fact that a signal for an actual acceleration a_act for said vehicle is noisy.

6. Method according to claim 5, where said force

equation appears as follows:

„ h (V_ref -V_act \

Fk+i =-^*{ ; a_act) + Fk, where - F_k+1 is a force which will impact on said vehicle at the next iteration;

- h is a discretion factor;

- γ is a calibration time;

- m is said mass of said vehicle;

v_ref is said reference speed;

- v_act is said actual speed;

- τ is a time parameter;

^{~ a}act is ^an actual acceleration for said vehicle; and

- F_k is a current force which impacts on said vehicle.

7. Method according to claim 6, where said control is made oscillation-free through the fact that said time

parameter τ is at least four times larger than said calibration time γ, τ≥ 4 * γ.

8. Method according to any of claims 2-7, where said model comprises a time parameter τ with a size which determines an appearance of a transient for said actual speed v_act toward said reference speed v_rej .

9. Method according to claim 8, where said size of said time parameter τ is related to a driving mode for said vehicle.

10. Method according to any of claims 2-9, where a transient for said actual speed v_act toward said reference speed v_ref is used to determine said mass m of said vehicle

11. Method according to claim 1, where said system is an engine in a vehicle, said at least one state consists of an actual engine speed <j⁾ _act for said engine, said reference value constitutes a reference engine speed d_ref for said engine, and said inertia is based on a moment of inertia / for said engine.

12. Method according to claim 11, where said model takes into consideration a difference between a change of an actual engine speed w_act for said engine and a change of a reference engine speed w_re/ f°^r said engine, where said difference depends on a time parameter τ.

13. Method according to any of claims 11-12, where said force equation has the following appearance:

^Tk+i =J*{ ; ^Mact) + ^Tk, where

- T_fc₊i is a torque which will be emitted by said engine at the next iteration;

- / is said moment of inertia for said engine;

- <j⁾ _ref is said reference engine speed;

^{~ ω}αα is said actual engine speed;

- τ is a time parameter;

- <j_a ^' _ct is a change of said actual engine speed; and

- T_k is a current torque emitted by said engine.

14. Method according to any of patent claims 11-13, where said model comprises a time parameter τ with a size which determines an appearance of a transient for said actual engine speed ω_αε(: toward said reference engine speed (j_ref .

15. Method according to patent claim 1, where said system is a temperature control system, said at least one state consists of an actual temperature T_act for a limited volume, said reference value consists of a reference temperature T_ref for said volume, and said inertia is based on a thermal inertia for said volume.

16. Method according to patent claim 15, where said model takes into consideration a difference between a change of an actual temperature T_act for said volume and a change of a reference temperature T_ref for said volume, where said

difference depends on a time parameter τ .

17. Method according to any of patent claims 15-16, where said equation has the following appearance:

P_k+1 + P_k, where

- ifc+i is a thermal effect which will be emitted in said volume at the next iteration;

- K is said thermal inertia;

- Tref is said reference temperature;

- T_act is said actual temperature;

τ is a time parameter;

^_ Tact is ^a change of said actual temperature; and

- P_fc is a current thermal effect emitted in said volume.

18. Method according to any of claims 15-17, where said model comprises a time parameter τ with a size which determines an appearance of a transient for said actual temperature T_act toward said reference temperature T_re · 19. Method according to claim 1, where said system is a system for acceleration limitation in a vehicle, said at least one state consists of an actual acceleration a_act for said vehicle, said reference value consists of a reference

acceleration _ref for said vehicle, and said inertia is based on a mass m of said vehicle.

20. Method according to claim 19, where said model takes into consideration a difference between said actual

acceleration a_act and said reference acceleration _rej .

21. Method according to any of claims 19-20, where said force equation has the following appearance:

Pk+i = m* (a_ref - a_act) + F_k , where

- F_k+1 is a force which will impact on said vehicle at the next iteration;

- m is said mass of said vehicle;

— &rgf is said reference acceleration;

- a_act is said actual acceleration; and

- F_k is a previous current force which impacts on said vehicle.

22. Method according to claim 1, where said system is a system for braking of a vehicle, said at least one state consists of an actual speed v_act for said vehicle, said

reference value consists of a maximum speed v_max for said vehicle, and said inertia is based on a mass m of said

vehicle .

23. Method according to claim 22, where said model takes into consideration a difference between an actual acceleration ^aact f°^r said vehicle and a reference acceleration a_re^ for said vehicle, where said difference depends on a time parameter τ .

24. Method according to any of claims 22-23, where said force equation has the following appearance:

^Bk+i =m* (^¾nx _T ^"¾t - a_act) + B_k , where

- B_k+1 is a force which will impact on said vehicle at the next iteration; - m is said mass m of said vehicle;

- v_rej is said reference speed;

- v_act is said actual speed;

- τ is a time parameter;

^{~ a}act is ^an actual acceleration for said vehicle; - B_k is a current braking force which impacts on said vehicle.

25. Computer program comprising a program code which, when said program code is executed in a computer, achieves that said computer carries out the method according to any of claims 1-24.

26. Computer program product comprising a computer readable medium and a computer program according to claim 25, where said computer program is comprised in said computer readable medium.

27. Control system for the control of at least one in a system toward a reference value, where said state

comprises an inertia and said control system is arranged carry out said control with the use of a control signal

characterised by

- a model device, arranged to provide a model comprising a force equation related to said system, where said control is model based; and

- a control device, arranged to provide said control signal so that a size of said control signal is proportionate to a change in said at least one state.

28. Control system according to claim 27, where said system is a cruise control system in a vehicle, said at least one state consists of an actual speed v_act for said vehicle, said reference value consists of a reference speed v_ref for said vehicle, and said inertia is related to a mass m of said vehicle .

29. Control system according to claim 27, where said system is an engine in a vehicle, said at least one state consists of an actual engine speed <j)_act for said engine, said reference value consists of a reference engine speed <j)_ref for said engine, and said inertia is based on a moment of inertia / for said engine.

30. Control system according to claim 27, where said system is a temperature control system, said at least one state consists of an actual temperature T_act for a limited volume, said reference value consists of a reference

temperature T_ref for said volume, and said inertia is based on a thermal inertia for said volume.

31. Control system according to claim 27, where said system is a system for acceleration limitation in a vehicle, said at least one state consists of an actual acceleration a_act for said vehicle, said reference value consists of a reference acceleration a_re^ for said vehicle, and said inertia is based on a mass m of said vehicle.

32. Control system according to claim 27, where said system is a system for braking of said vehicle, said at least one state consists of an actual speed v_act for said vehicle, said reference value consists of a maximum speed v_max for said vehicle, and said inertia is based on a mass m of said

vehicle .